Data anonymization to balance privacy and utility of online social media network data

Abstract

There are tremendous users of social media networks, and they keep growing day by day. In recent years, social media networks have sparked widespread interest among the general public because they offer a simple and appealing form of communication. Users’ data is at risk due to their expanding contact with social networks, necessitating securing their privacy. Users communicate with each other on social media and share information that is vital and private. This user’s information is on the tip of attraction, as many third-party ethical users use it for good causes like increasing the new customers by analyzing their needs. Still, unethical users misuse it for destructive purposes like stealing data, burglary, or personification. And hence the data owner’s task is to preserve the privacy of information. Before publishing such information in the public domain, it needs to be anonymized so that a third party cannot misuse it. Again balancing privacy and utility also need to be focused. The proposed improved Mondrian algorithm anonymizes data by partitioning and enhances privacy and utility balance on the Adult dataset same as with measures PIRL, NCP and DP.

Subject Classification:

• 68 Computer science)
• 68T01 (General topics in artificial intelligence)

Keywords:

• Anonymization
• Quasi-identifiers
• Identifier
• Sensitive attribute
• k-anonymity
• Mondrian
• Privacy
• Utility